Coready AI
Coready AI is an assistant built into your admin panel. Ask it about your catalog, orders and contractors, have it look things up, or describe a change in plain words — "make the blue mug 12.50", "add a red mug for 9 under SKU MUG-2" — and it prepares the change for you to confirm. Open it with Coready AI in the top bar, next to what you have left to spend on it.
You do not need an account with an AI company or a key of your own: Coready runs the models, and each answer costs what it used — see What an answer costs.
Conversations
Every conversation is saved, so you can come back to it later. They are yours alone — nobody else in your company sees them — and they belong to the company you were working in when you started them.
The ⋯ menu next to a conversation renames, archives or deletes it:
- Archive moves it out of the list into Archived. An archived conversation can be read but not continued; Restore brings it back. A change that was still waiting for your confirmation is declined when you archive.
- Delete removes its messages for good; what it cost stays in your usage history.
A very long conversation eventually asks you to start a new one. Start a new one whenever the topic changes, too: answers are faster and cheaper when the assistant has less to read.
Choosing a model
The menu at the top of a conversation picks the model that writes the next answer:
| Fast | Quick answers and lookups. The cheapest. |
| Balanced | Good for most work: questions, several steps, drafting changes. |
| Most capable | The hardest questions and long answers. Costs the most per answer. |
You can switch at any point in a conversation.
Images
Add up to four images to a message with the image button, by pasting them, or by dropping them onto the message box: screenshots of a supplier's price list, a product photo, a table from a spreadsheet. Large images are scaled down before they are sent. They stay with the conversation, visible only to you.
Prices from a screenshot
Send a screenshot of a price table and say which product and option it is for — "apply this table to the Brazil Cerrado, option Bag size". Coready AI reads the table, matches its rows to the option's values on the product (a "250g bag" in the image is your "250 g"), and shows every value and quantity with its old and new price. Quantity steps the table no longer has are shown as removed. Nothing is saved until you confirm.
What it can do
- Read: products (with their options, quantity prices and price types), categories, manufacturers, options and their values, orders and their lines, and contractors — your customers and suppliers.
- Explain Coready: it reads these help pages, the documentation of the modules your company can use, and what you have in Coready Studio — whether your plan includes it and your websites.
- Prepare changes: create or change a product, set an option's quantity prices from a price table, create or change a category, a manufacturer or a contractor, create an option or add a value to one.
It cannot delete anything, create or change orders, or change settings, your team, billing or your website's pages. It tells you where in the panel to do those.
It sees only what you can see: the records of the company you are working in, with your own rights.
Nothing changes without you
When Coready AI wants to change something, it shows a card with exactly what will change — the old value crossed out, the new one beside it — and two buttons:
- Confirm saves the change, the same way as if you had made it on the record's page. It shows up in the record's history under your name.
- Decline leaves everything as it was.
If you write a new message instead of answering, the change is declined. A new product is created hidden from your website unless you asked for it to be shown, so you can look it over first.
What an answer costs
An answer costs exactly what it used: the AI provider's list price for the tokens it read and wrote, plus 25%. Tokens are the pieces of text a model reads and writes — roughly three or four characters each; your message, the conversation so far, images and what the assistant looks up all count as tokens read.
price = (tokens read × input price + tokens written × output price) ÷ 1,000,000
Prices per million tokens, with the 25% included, in US dollars:
| Model | Input | Output |
|---|---|---|
| Claude Sonnet 5.5 | $2.50 | $12.50 |
| Claude Opus 5.5 | $5.00 | $25.00 |
| Claude Haiku 5.5 | $0.125 | $0.625 |
| GPT-6.1 Sol | $2.50 | $12.50 |
| GPT-6 Luna | $0.125 | $0.625 |
| Gemini 3.8 Flash | $1.875 | $9.375 |
| Gemini 2.5 Flash | $0.375 | $3.125 |
If your balance is in euros, the dollar price is converted at the European Central Bank's reference rate of that day. Account → AI usage shows today's prices in your currency.
An example. You ask a question with some context — 3,000 tokens read — and get a short answer of 600 tokens on Claude Sonnet 5.5: 3,000 × $2.50 + 600 × $12.50 = $15,000 per million = $0.015, about one and a half cents. A long answer or the most capable model costs more; a quick lookup on a fast model costs a fraction of a cent.
How it is paid. Your plan includes some Coready AI every month for the whole company; it is used first. After it, answers are paid from your company's balance (see Billing). Prices are without tax: tax is added once, on the invoice when the balance is topped up.
- After each answer the conversation shows what it cost and what is left.
- Account → AI usage lists every answer with its model, its tokens and its cost, what the company has spent this month, and what is left of the included AI.
- Nothing left? Coready AI says so and charges nothing more. An admin tops up the balance; a member presses Ask an admin.
- A personal limit. An admin can give each person a monthly Coready AI limit. Someone who reaches it asks for a higher one the same way.
If an answer fails part-way — the AI company is busy, or you press Stop — you pay only for what was used up to that point. Up to 60 answers an hour per person.